I don't understand what the motivation is for a change like this? What's
the advantage over what exists today?
-RPM
On Thu, 2007-01-04 at 15:14 +0100, Anselm R. Garbe wrote:
> Hi there,
>
> I'm thinking about a different kind of tiled layout which I try
> to describe in the following...
>
> What I have in mind is something like a customizable grid layout
> which displays clients within a ring list.
>
> By default the grid might display 4 clients, but the grid size
> can be customized from 0..n^2 grids, which means 1, 2, 4, 9, 16,
> ..., 0 means arrange arbitrary clients in a grid.
>
> Next, there won't be a zoom() as we know it, instead this
> functionality could be merged into focus{next,prev}, which means
> the layout changes if the next/prev client is not visible, e.g.
>
> AAABBB
> AAABBB
> AAABBB
>
> (assume a grid layout with 2 clients being visible only)
>
> i) If A is focused and focusnext() is performed, B will be focused.
> ii) If A is focused and focusprev is performed, and there is a
> third client, the layout will look like:
>
> CCCAAA
> CCCAAA
> CCCAAA
>
> with C being focused. If there is no other client than A and B
> then the layout will look like
>
> BBBAAA
> BBBAAA
> BBBAAA
>
> with B being focused.
>
> All other cases should be easy to grasp. You see that this
> concept scales up to arbitrary clients.
>
> Because this is a ring list, there must be a way to swap the
> order of such tiled clients in both directions, hence there
> should be a swapnext() and swapprev() as replacements for
> zoom().
>
>
> What do people think about this concept? I know it is totally
> different from current tiled layout, but it seems worth to being
> checked out - it's not drastically more complicated than the
> current concept. The grid size could be grown/shrinked easily
> with two actions (like current grow-/shrink master area)...
>
> It also works well with mouse focus...
>
> Regards,
